Rescaling of cognitive flexibility inventory by criticism of Turkish adaptation form

Abstract

Cognitive flexibility is the evaluation of events from a broad perspective rather than evaluating them rigidly. The use of the Cognitive Flexibility Inventory (CFI) by adaptation to different cultures and the increase in citations in recent years indicate that this assessment instrument is up-to-date and practical. Therefore, it is necessary and important to investigate the psychometric properties of this frequently used assessment instrument with different techniques. In this study, it was aimed to rescale the CFI items and model parameters according to item response theory and inspect differentiation for CFI items according to gender. Research data were collected from a total of 750 undergraduate students (564 female and 186 male) via Google Forms. To decide on the best model for the cognitive flexibility construct, bifactor, multidimensional, and unidimensional item response theory models were tested according to item fit and global model-data fit statistics. Besides, the graded response model was used to calibrate the CFI items. As a result of IRT analysis, it was determined that the model with which the CFI fits best was the bifactor model. After IRT analysis, differentiation was inspected via different methods. After the differential item functioning analysis, expert opinions were obtained for items which had a potential bias. As a conclusion, there was a consensus only on the idea that item 12 showed gender bias. With 5-point Likert-type scaling, gender bias existed. A new adaptation for a 7-point Likert type should be implemented and gender bias should be examined.